> > I was more concerned with the larger question that contexts like this, > assignment of __init function pointers to dynamic data structure > members, is not checked. > > The checker has no way to make sure that pbm->scan_bus() is only > made in __init contexts, even though we can plainly see that is > the case here. > > What would be really clever would be a way to mark datastructure > members as "this is a pointer to an __init function".
Today with the section resolution used by modpost I can annotate the full datastructure to say that references to __*init/__*exit should not be flagged - thedata are simply put in a dedicated section. The more fine grained approach you are requesting needs assistance with sparse I think. I will explore that later when I get section mismatch stuff under control. With section mismatch warnings being converted to errors soon I foresee a bit nosie I need to concentrate on first. But at least sparc64 is in pretty good shape now. Sam - To unsubscribe from this list: send the line "unsubscribe sparclinux" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html